Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Tensorflow 2.x implementation of the DTLN real time speech denoising model. With TF-lite, ONNX and real-time audio processing support.
| Date | Stars |
|---|---|
| 2026-07-24 | 730 |
| 2026-07-25 | 731 |
| 2026-07-28 | 731 |
| 2026-07-30 | 731 |
| 2026-08-06 | 731 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Dual-signal Transformation LSTM Network
+ Tensorflow 2.x implementation of the stacked dual-signal transformation LSTM network (DTLN) for real-time noise suppression.
+ This repository provides the code for training, infering and serving the DTLN model in python. It also provides pretrained models in SavedModel, TF-lite and ONNX format, which can be used as baseline for your own projects. The model is able to run with real time audio on a RaspberryPi.
+ If you are doing cool things with this repo, tell me about it. I am always curious about what you are doing with this code or this models.
---
The DTLN model was handed in to the deep noise suppression challenge ([DNS-Challenge](https://github.com/microsoft/DNS-Challenge)) and the paper was presented at Interspeech 2020.
This approach combines a short-time Fourier transform (STFT) and a learned analysis and synthesis basis in a stacked-network approach with less than one million parameters. The model was trained on 500h of noisy speech provided by the challenge organizers. The network is capable of real-time processing (one frame in, one frame out) and reaches competitive results.
Combining these two types of signal transformations enables the DTLN to robustly extract information from magnitude spectra and incorporate phase information from the learned feature basis. The method shows state-of-the-art performance and outperforms the DNS-Challenge baseline by 0.24 points absolute in terms of the mean opinion score (MOS).
For more information see the [paper](https://www.isca-speech.org/archive/interspeech_2020/westhausen20_interspeech.html). The results of the DNS-Challenge are published [here](https://www.microsoft.com/en-us/research/academic-program/deep-noise-suppression-challenge-interspeech-2020/#!results). We reached a competitive 8th place out of 17 teams in the real time track.
---
For baseline usage and to reproduce the processing used for the paper run:
```bash
$ python run_evaluation.py -i in/folder/with/wav -o target/folder/processed/files -m ./pretrained_model/model.h5
```
---
The pretrained DTLN-aec (the DTLN applied to acoustic echo cancellation) can be found in the [DTLN-aec repository](https://github.com/breizhn/DTLN-aec).
---
Author: Nils L. Westhausen ([Communication Acoustics](https://uol.de/en/kommunikationsakustik) , Carl von Ossietzky University, Oldenburg, Germany)
This code is licensed under the terms of the MIT license.
---
### Citing:
If you are using the DTLN model, please cite:
```BibTex
@inproceedings{Westhausen2020,
author={Nils L. Westhausen and Bernd T. Meyer},
title={{Dual-Signal Transformation LSTM Network for Real-Time Noise Suppression}},
year=2020,
booktitle={Proc. Interspeech 2020},
pages={2477--2481},
doi={10.21437/Interspeech.2020-2631},
url={http://dx.doi.org/10.21437/Interspeech.2020-2631}
}
```
---
### Contents of the README:
* [Results](#results)
* [Execution Times](#execution-times)
* [Audio Samples](#audio-samples)
* [Contents of the repository](#contents-of-the-repository)
* [Python dependencies](#python-dependencies)
* [Training data preparation](#training-data-preparation)
* [Run a training of the DTLN model](#run-a-training-of-the-dtln-model)
* [Measuring the execution time of the DTLN model with the SavedModel format](#measuring-the-execution-time-of-the-dtln-model-with-the-savedmodel-format)
* [Real time processing with the SavedModel format](#real-time-processing-with-the-savedmodel-format)
* [Real time processing with tf-lite](#real-time-processing-with-tf-lite)
* [Real time audio with sounddevice and tf-lite](#real-time-audio-with-sounddevice-and-tf-lite)
* [Model conversion and real time processing with ONNX](#model-conversion-and-real-time-processing-with-onnx)
---
### Results:
Results on the DNS-Challenge non reverberant test set:
Model | PESQ [mos] | STOI [%] | SI-SDR [dB] | TF version
--- | --- | --- | --- | ---
unprocessed | 2.45 | 91.52 | 9.07 |
NsNet (Baseline) | 2.70 | 90.56 |Excerpt of 16,076 characters
Read on GitHub
100
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:e818d6962a613644, topic:deep-learning, topic:tensorflow
matched fp:e818d6962a613644, topic:onnx
matched fp:e818d6962a613644, topic:audio-processing